Modeling photoinduced toxicity of PAHs based on DFT-calculated descriptors

Chemosphere 76 (2009) 999–1005

Contents lists available at ScienceDirect

Chemosphere journal homepage: www.elsevier.com/locate/chemosphere

Modeling photoinduced toxicity of PAHs based on DFT-calculated descriptors Ying Wang a, Jingwen Chen a,*, Fei Li a, Hong Qin a, Xianliang Qiao a, Ce Hao b a

Key Laboratory of Industrial Ecology and Environmental Engineering (MOE), Department of Environmental Science and Technology, Dalian University of Technology, Linggong Road 2, Dalian 116024, PR China Carbon Research Laboratory, Center for Nano Materials and Science, School of Chemical Engineering, State Key Laboratory of Fine Chemicals, Dalian University of Technology, Zhongshan Road 158, Dalian 116012, China

b

a r t i c l e

i n f o

Article history: Received 31 December 2008 Received in revised form 30 March 2009 Accepted 3 April 2009 Available online 8 May 2009 Keywords: QSAR Density functional theory Photoinduced toxicity Mechanism Polycyclic aromatic hydrocarbons

a b s t r a c t Quantitative structure–activity relationships (QSARs) were established for photoinduced toxicity of polycyclic aromatic hydrocarbons (PAHs) to two aquatic species. Partial least squares (PLS) regression and molecular structural parameters calculated by density functional theory (DFT) were employed for model development. Two QSAR models were established and their high R2 and Q 2CUM values indicated their good goodness-of-fit, robustness and internal predictive power. The descriptors that describe the partition behavior, light absorbance, and generation of reactive free radicals were found to be successful in modeling the photoinduced toxicity. The average molecular polarizability (a), energy gap (EGAP) between the energy of the lowest unoccupied molecular orbital and the highest occupied molecular orbital, lowest triplet excitation energy (ET1) and vertical electron affinity at the lowest excited triplet (VEAT1) were the main molecular structural factors. Polarizability which determines the partition of PAHs between water and lipid governs the photoinduced toxicity of selected PAHs. Moreover, the photoinduced toxicity increased with the decreasing of EGAP probably due to better spectral overlap. The parameter, VEAT1 that characterizes the ability of PAH anion radical (PAH) generation from excited triplet state PAH (3PAH*), is also related with the photoinduced toxicity. This investigation will make us gain more insight into the photoinduced toxicity mechanism and assess the applicability of various DFT-based descriptors to toxicological QSARs. Ó 2009 Elsevier Ltd. All rights reserved.

1. Introduction Upon co-exposure to light a number of organic compounds such as polycyclic aromatic compounds (Oris and Giesy, 1985; Newsted and Giesy, 1987; Huang et al., 1995; Arfsten et al., 1996; Wiegman et al., 1999; Lampi et al., 2005), chlorophenols (Svenson and Hynning, 1997) and nitrotoluenes (Johnson et al., 1994) can exert photoinduced toxicity to aquatic species. Polycyclic aromatic hydrocarbons (PAHs) are a class of ubiquitous pollutants in the environment and known for their photoinduced toxicity, which occurs through two types of mechanisms, photomodification to more toxic photoproducts (Huang et al., 1993; McConkey et al., 1997), and photosensitization reactions with the formation of reactive oxygen species (ROS) (Little et al., 2000; Huovinen et al., 2001). For the photoinduced toxicity of PAHs, photosensitization mechanism has been considered as being dominant, although a bipartite mechanism has been proposed (El-Alawi et al., 2002; Lampi et al., 2007). Due to the ever-increasing and great number of organic pollutants in the aquatic environment, it is of importance to develop * Corresponding author. Tel./fax: +86 411 8470 6269. E-mail address: [email protected] (J. Chen). 0045-6535/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2009.04.010

models that relate molecular structure characters with the toxicological activities. Quantitative structure–activity relationships (QSARs) studies are helpful to understand and predict the photoinduced toxicity. Based on the bipartite mechanism, photosensitization factor (PSF) and photomodification factor (PMF) have been proposed and successfully employed to establish QSAR models (Huang et al., 1997; El-Alawi et al., 2002; Lampi et al., 2007). However, quite a few empirical parameters were needed to obtain PSF and PMF. It would be of interests and significance to develop photoinduced toxicity QSAR models employing theoretical molecular structural parameters. Such QSAR models are also favorable to prior predictions. Some previous studies found that the energy gap (EGAP) between the energy of the lowest unoccupied molecular orbital (ELUMO) and the highest occupied molecular orbital (EHOMO) were important in explaining the photoinduced toxicity (Mekenyan et al., 1994; Veith et al., 1995a, b; Dong et al., 2002; Grote et al., 2005). In these studies, ELUMO and EHOMO were all calculated by semiempirical quantum chemical methods. In recent years, parameters derived from density functional theory (DFT) have been proved to be attractive to predict biological activities because of their better accuracy (Enoch et al., 2008; Pasha et al., 2008). Correlations between DFT-based excited states energy and photoinduced

1000

Y. Wang et al. / Chemosphere 76 (2009) 999–1005

toxicity have also been reported (Betowski et al., 2002; Wang et al., 2007). However, there is a lack of quantitative investigation between the photoinduced toxicity and DFT descriptors. The purpose of this study was to develop QSAR models on photoinduced toxicity of PAHs with DFT descriptors. We analyzed the processes that may contribute to photoinduced toxicity, such as bioaccumulation of compounds in the biological tissue, light absorption, and generation of free radicals and the molecular structural parameters that are relevant with processes were derived by DFT. The constructed QSAR models were applied to further interpret the photoinduced toxicity mechanism. 2. Materials and methods 2.1. Data set The photoinduced toxicity data set of 14 PAHs to Daphnia magna and 12 PAHs to Scenedesmus vacuolatus were adopted from Lampi et al. (2005) and Grote et al. (2005). The logarithm values of medium effect concentration (EC50) in the unit of nM were used as the toxic endpoint of the QSAR study. Molecular structures and the sequence numbers for the 19 PAHs included in this study are shown in Fig. 1. 2.2. Descriptors selection The internal concentration of the target compound in an organism can provide a better basis for assessing the toxic effect (Escher and Hermens, 2002; Grote et al., 2005). The partition process of chemicals between water and lipid should be characterized by intermolecular forces, in which the dispersion force should play a key role during the partition process. Moreover, molecular volume determines the ability of passing the biomembrane channel. Thus molecular molar volume (V) and average molecular polarizability

(a) were selected as descriptors. Some descriptors describing the photochemical properties such as ELUMO, EHOMO and EGAP were also adopted. They can provide the information of light absorbance of compounds (Mekenyan et al., 1994; Chen et al., 1996). The relative great contribution of photosensitization to photoinduced toxicity of PAHs was pointed out compared with that of photomodification (Dong et al., 2000; Lampi et al., 2007). Fig. 2 summarized the potential photosensitization type I and type II reactions mechanism of PAHs (Foote, 1991; Fasnacht and Blough, 2003; Toyooka and Ibuki, 2007). It has been commonly believed that the photosensitization reactions proceed at the excited triplet state because of its relative long life. In a type I reaction, excited triplet state PAH (3PAH*) can be reduced to highly reactive PAH anion radical (PAH) by biological substrates such as amino acids and DNA/RNA bases. Then PAH is capable of reacting with substrates or transferring its electron to molecular oxygen, leading to the generation of superoxide anion  (O 2 ). O2 can induce the generation of other ROS such as hydrogen peroxide (H2O2) and hydroxyl radicals (OH) through dismutation reaction and Fenton reaction. The vertical electron affinity at the lowest excited triplet (VEAT1) is a measure of the energy gain upon reduction from the T1 state; the larger the VEAT1, the easier it is for the compound to oxidize a nearby species. VEAT1 can describe the ability of the PAH generation from 3PAH*. Furthermore, the vertical electron affinity at singlet ground state (VEAS0) can reflect the  and O 2 formation capability through the reaction between PAH molecular oxygen. The PAH cation radical (PAH+) was involved in the mechanism of PAH photodegradation and it can also induce the formation of  O 2 (Fasnacht and Blough, 2003). To investigate the O2 production in this process, we choose vertical ionization energy at excited triplet state (VIET1) to describe the reducing ability of 3PAH*. The smaller the VIET1, the easier for the compound it is to reduce the oxygen molecular.

(1)

(2)

(3)

(4)

(5)

anthracene

benzo[a]pyrene

benzo[a]anthracene

benzo[b]anthracene

benzo[b]fluorene

(6) fluorene

(7) fluoranthene

(8) benzo[ghi]perylene

(9) dibenz[a,h]anthracene

(10) dibenzo[a,i]pyrene

(11) chrysene

(12) phenanthrene

(13) indeno[1,2,3-cd]pyrene

(14) benzo[k]fluoranthene

(15) benzo[b]fluoranthene

(16) pyrene

(17) benzo[e]pyrene

(18) benzo[ghi]fluoranthene

(19) 2-phenylnaphthalene

Fig. 1. Molecular structures of 19 PAHs under study.

Y. Wang et al. / Chemosphere 76 (2009) 999–1005

1001

calculations were performed on the optimized geometries to ensure the systems had no imaginary vibration frequencies. The polarizability was calculated according to the frequency analysis results. Single point calculations were made for neutral molecules, anions and cations at their respective optimized neutral geometries. Excited singlet and triplet energies were calculated using time dependent-density functional theory (TD-DFT). The solvent effects were included through singlet point and TD-DFT calculations on the optimized geometries at B3LYP/6-31 + G(d,p) level, by employing the self-consistent reaction field (SCRF) method with the integral equation of the polarized continuum model (IEFPCM). Water was used as solvent and the diffusion functions are essential for the treatment of anion and cation radicals. 2.4. Partial least squares (PLS) analysis Fig. 2. Type I and type II mechanisms of PAHs photosensitization. This is adapted based on the previous studies (Foote, 1991; Fasnacht and Blough, 2003; Toyooka and Ibuki, 2007). 1PAH represents a ground state PAH; 1PAH*/3PAH* represents 1 singlet/triplet excited state PAH; hm, photon; O 2 , superoxide anion; O2 , singlet oxygen.

In a type II reaction, the excitation energy of 3PAH* can transfer its energy to the surrounding molecular oxygen to generate singlet oxygen (1 O2 ). Thus it is also necessary to adopt lowest triplet excitation energy (ET1) to investigate the possibility of transfer energy. In addition, E2T1 and E2GAP were also selected as predictor variables due to their significance as indicated by the previous studies (Newsted and Giesy, 1987; Mekenyan et al., 1994). Thus, the DFT descriptors listed in Table 1 can be classified into three groups: (1) partition descriptors including V and a, (2) light absorbance descriptors including ELUMO, EHOMO and EGAP, (3) free radicals generation related descriptors including ET1, VEAT1, VEAS0 and VIET1. Table 2 shows the values of the calculated DFT-based descriptors for selected PAHs. Average molecular polarizability (Karelson et al., 1996) (a) is calculated as

a¼

axx þ ayy þ azz 3

PLS regression analysis was performed using the Simca-S package (Version 6.0, Umetri AB & Erisoft AB) because it can analyze data with strongly collinear, noisy and numerous predictor variables. Simca-S adopts leave-many-out cross validation to determine the number of significant PLS components (A). Crossvalidation simulates how well a model predicts new data, and gives a statistical Q 2CUM (the fraction of the total variation of the dependent variables that can be predicted by all the extracted components) for the final PLS model. When Q 2CUM is larger than 0.5, the statistical models believed to have a good internal predictive power and robustness (Niu et al., 2006). If irrelevant or redundant descriptors are included, the internal predictive power and robustness of the PLS model may decrease, and the interpretation of the model will become difficult. So it is necessary to select important predictor variables. The variable selection procedure described previously (Ding et al., 2005), which is based on Q 2CUM and variable importance in the projection (VIP), was adopted in the current study. VIP indicates the relevance of predictor variables in explaining the dependent variables, and predictor variables with large VIP are the most relevant for explaining response variable.

where axx, ayy, and azz are the diagonal elements in the standard orientation of molecular polarizability tensor.

3. Results and discussion

2.3. Quantum chemical calculations

3.1. PLS regression for predicting photoinduced toxicity

All the DFT computations were performed by Gaussian 03 programs (Frisch et al., 2003). Initial geometries of PAHs, their neutral

The following QSAR models were developed using DFT-based descriptors:

Daphnia magna : log EC 50 ¼ 9:721  103 a þ 9:017  102 E2GAP þ 1:335  101 E2T1  1:849VEAT1 þ 9:976 n ¼ 14;

A ¼ 2;

Scenedemus v acuolatus : log EC 50 ¼

Q 2CUM ¼ 0:738;

R2 ¼ 0:820;

RMSE ¼ 0:502;

6:396  10 þ 9:463  102 E2GAP  þ 1:118  101 E2T1 þ 4:443  101 ET1 3

a

p < 0:001

ð1Þ

1:087VEAT1

þ 1:161  101 EHOMO þ 5:819 n ¼ 12;

A ¼ 2;

Q 2CUM ¼ 0:862;

R2 ¼ 0:921;

molecular, radical anion and radical cation forms of compounds were pre-optimized by semiempirical method PM3 Hamiltonian and were then optimized at the B3LYP/6-31G(d,p) level. Frequency

RMSE ¼ 0:252;

p < 0:001

ð2Þ

where n stands for the number of PAHs, A is the number of PLS components, R2 is coefficient of determination and RMSE is root mean squared error. The predictor variables in QSAR models are ar-

1002

Y. Wang et al. / Chemosphere 76 (2009) 999–1005

Table 1 Molecular structural descriptors. No. 1 2 3 4 5 6 7 8 9 10

Descriptors

Definition

Units

V

Molecular molar volume (electron density volume) Average molecular polarizability Energy of the lowest unoccupied molecular orbit Energy of the highest occupied molecular orbit ELUMO  EHOMO Vertical electron affinity at singlet ground state Vertical ionization energy at singlet ground state Lowest triplet excitation energy VEAS0  ET1 VIES0  ET1

cm3/mol atom unit eV eV eV eV eV eV eV eV

a EHOMO ELUMO EGAP VEAS0 VIES0 ET1 VEAT1 VIET1

Table 2 Values of observed log EC50 (nM) and selected DFT descriptors for PAHs. No.

Compounds

Model (1) log EC50

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Anthracene Benzo[a]anthracene Benzo[a]pyrene Benzo[b]anthracene Benzo[b]fluorene Benzo[e]pyrene Benzo[ghi]perylene Chrysene Dibenz[a,h]anthracene Dibenzo[a,i]pyrene Fluoranthene Fluorene Phenanthrene Pyrene Benzo[b]fluoranthene Benzo[k]fluoranthene Indeno[1,2,3-cd]pyrene Benzo[ghi]fluoranthene 2-phenylnaphthalene

1.80 0.62 0.59 1.01 1.61 0.11 0.32 1.24 0.30 0.36 1.29 4.23 3.43 1.36

a b

a

Model (2)log EC50

b

2.02 1.06 0.40

0.86

1.42 3.58 1.54 0.80 1.12 0.36 1.01 2.56

V

a

ELUMO

EHOMO

EGAP

ET1

VEAS0

VEAT1

VIET1

153 158 180 190 166 172 202 159 229 231 153 138 116 163 167 159 220 153 163

229 306 360 334 267 328 379 294 388 466 251 184 214 263 385 348 393 287 244

1.96 1.88 2.06 2.38 1.58 1.76 2.02 1.61 1.82 2.11 2.12 1.13 1.35 1.81 2.08 2.06 2.34 2.20 1.50

5.51 5.61 5.40 5.13 5.82 5.71 5.50 5.82 5.67 5.33 6.07 6.07 6.03 5.61 6.02 5.67 5.64 6.20 5.95

3.55 3.73 3.34 2.75 4.24 3.95 3.48 4.21 3.85 3.22 3.95 4.94 4.68 3.80 3.94 3.61 3.30 4.00 4.45

1.83 2.07 1.81 1.15 2.54 2.34 2.03 2.52 2.26 1.74 2.37 3.10 2.76 2.14 2.44 2.23 1.92 2.53 2.64

2.15 2.09 2.25 2.58 1.78 1.96 2.21 1.80 2.02 2.31 2.32 1.34 1.54 2.01 2.28 2.26 2.54 2.40 1.70

3.98 4.16 4.07 3.73 4.31 4.30 4.25 4.32 4.28 4.05 4.69 4.44 4.31 4.15 4.72 4.49 4.45 4.93 4.33

3.51 3.36 3.40 3.79 3.12 3.19 3.29 3.12 3.23 3.40 3.53 2.82 3.10 3.30 3.41 3.26 3.55 3.50 3.16

Data from Lampi et al. (2005). Data from Grote et al. (2005).

ranged in the order of decreasing VIP values. The QSAR models have good goodness-of-fit, robustness and internal predictive power according to the high statistical significance. As shown in Fig. 3, the observed log EC50 values are in good agreement with the predicted values. The VIP values and PLS weights for the two models are listed in Table 3. The VIP values in the two models for a are the largest indicating that a governs the photoinduced toxicity of PAHs more significantly than the other predictors. The factors governing log EC50 can be interpreted by the PLS weights of the variables. The information carried by the molecular structural descriptors was con-

4.0

(a) Daphnia magna

(b) Scenedemus vacuolatus

4

logEC50 (predicted)

logEC50 (predicted)

5

densed into two latent variables (Table 3). The first PLS component is primarily loaded on a, E2GAP and E2T1 ; suggesting that it mostly condenses the information related with intermolecular interactions and excited state properties. The second PLS component, which is mainly loaded on VEAT1, reflects the information on the generation of free radical. The negative PLS weights and coefficients of a in the two models indicate the negative correlation relationship between a and log EC50. Photoinduced toxicity of studied PAHs increases with increasing a. As indicated by the pseudo-regression coefficients for the predicted variables, the increasing EGAP and ET1 values of studied PAHs results in the

12

3

13

2

51 814

1 0

7

6 2 11 4 93

R 2 = 0.820 Q 2CUM = 0.738 RMSE = 0.502

10

-1 -1

0

1 2 3 4 logEC50 (observed)

5

13

3.2 19

2.4 1.6

2 18

14 11

3

0.0 0.0

17

0.8

2

R = 0.921 Q 2CUM = 0.862 RMSE = 0.252

16 15 7

0.8

1

1.6

2.4

3.2

4.0

logEC50 (observed)

Fig. 3. (a) Predicted toxicity (log EC50) of PAHs versus their observed toxicity to Daphnia magna. (b) Predicted toxicity (log EC50) of PAHs versus their observed toxicity to Scenedesmus vacuolatus. The sequence numbers of PAHs correspond with those in Fig. 1.

1003

Y. Wang et al. / Chemosphere 76 (2009) 999–1005 Table 3 VIP values and PLS weights for Models (1) and (2). Models

Variables

VIP

W*[1]

W*[2]

Model (1)

a

1.155 0.996 0.929 0.901

0.609 0.566 0.520 0.197

0.613 0.117 0.032 0.823

1.231 1.168 0.934 0.915 0.909 0.762

0.541 0.546 0.420 0.122 0.403 0.235

0.565 0.253 0.049 0.782 0.086 0.387

E2GAP E2T1 VEAT1

a

Model (2)

E2GAP E2T1 VEAT1 ET1 EHOMO

decrease of photoinduced toxicity. Moreover, there are parabolic relationships between the photoinduced toxicity of PAHs and EGAP/ET1 according to the two QSAR equations. The results of the present study were consistent with the developed curve-linear (between photoinduced toxicity and EGAP/ET1) models in previous studies (Newsted and Giesy, 1987; Mekenyan et al., 1994). The negative PLS weights and coefficients of VEAT1 in the two models demonstrate the negative correlation relationship between VEAT1 and log EC50. 3.2. Mechanistic interpretation The two developed QSAR models were applied to further interpret the mechanism of photoinduced toxicity to aquatic species of PAHs as follows: It is interpretable that a is the most important influential molecular structure parameter for the photoinduced toxicity of PAHs. PAHs are non-polar or weak-polar and they tend to be par-

Table 4 Singlet excitation energies (E) and oscillator strengths (f) of 19 PAHs

a

titioned into nonpolar environments such as the phospholipids in cell membranes (Newsted and Giesy, 1987) and the partition is governed by the intermolecular force. The dispersive force that can be described by a is the most important intermolecular force and is able to reflect the partition behavior between PAHs and phospholipids. The stronger the dispersive force, the more easily PAHs partition in nonpolar phospholipids. Thus, PAHs with relative large a may more easily concentrate in organisms to exert their toxicity. Moreover, it can be found that a, the deformability of electron cloud, correlates well with molecular volume (n = 19, r = 0.851, p < 0.001). However, chemicals with the overlarge molecular volume may hardly cross the cell membrane, leading to lower toxicity. Among the studied PAHs, V of dibenzo[a,i]pyrene (#10) is the largest (231 cm3/mol). Thus, the photoinduced toxicity of dibenzo[a,i]pyrene is over-predicted (Fig. 3a). Some theoretical studies were performed successfully using TDDFT to explore the phototoxic reactions of some chemicals (Llano et al., 2003; Musa and Eriksson, 2007; Shen et al., 2007). In the present study, TD-DFT was applied to calculate the excited state parameters of PAHs. The potency of the photoinduced toxicity of studied PAHs is also influenced by their inherent photochemical properties such as EGAP and ET1. EGAP is a parameter characterizing the molecular electronic structure related to light absorption (Mekenyan et al., 1994; Chen et al., 1996; Wang et al., 2007). As we know, the wavelength of environmental solar radiation that reaches the earth surface is greater than 290 nm. PAHs with small EGAP are ‘‘red-shifted” because of the larger conjugated system (Newsted and Giesy, 1987). Table 4 shows the calculated single excitation energies and oscillator strengths calculated by TD-DFT. Singlet excitation energies of a compound reflect its ultravioletvisible absorption wavelengths. Oscillator strength is a measure of transition intensity and is related to the molar absorption

in water.

1

E (eV) k (nm) f

3.19 388.9 0.079

3.85 321.9 0

4.51 275.1 0

4.81 257.9 0

4.90 252.8 2.253

11

E (eV) k (nm) f

3.37 368.1 0.006

3.52 352.6 0.255

4.13 300.1 0.049

4.57 271.5 0.242

4.71 263.3 0.077

2

E (eV) k (nm) f

3.31 374.1 0.076

3.53 351.1 0.003

4.13 300.5 1.054

4.372 283.8 0

4.53 273.9 0.155

12

E (eV) k (nm) f

4.46 277.7 0.372

4.65 266.6 0.219

4.82 257.1 0.013

5.43 228.3 0.004

5.49 225.9 0.032

3

E (eV) k (nm) f

3.08 402.3 0.384

3.40 365.0 0.019

3.98 311.4 0.145

4.14 299.7 0.521

4.25 291.5 0.001

13

E (eV) k (nm) f

3.95 314.0 0.004

4.18 296.9 0.121

4.67 265.8 0.162

4.79 259.1 0.907

5.04 245.8 0.033

4

E (eV) k (nm) f

2.42 512.0 0.067

3.47 357.2 0

3.61 343.1 0

3.89 319.0 0

4.39 282.3 3.001

14

E (eV) k (nm) f

3.60 344.7 0.401

3.73 332.2 0.001

4.27 290.4 0

4.51 275.0 0.383

4.62 268.5 0

5

E (eV) k (nm) f

3.86 321.2 0.301

4.00 310.1 0.022

4.55 272.5 0.417

4.64 267.1 0.662

4.89 253.8 0.045

15

E (eV) k (nm) f

3.43 361.0 0.038

3.50 354.2 0.343

3.65 340.1 0.038

4.14 299.7 0.157

4.30 288.1 0.249

6

E (eV) k (nm) f

3.61 343.7 0.001

3.65 339.9 0.271

4.07 304.6 0.298

4.23 292.9 0.211

4.35 285.1 0.015

16

E (eV) k (nm) f

3.17 390.8 0.351

3.51 353.2 0.001

3.70 335.2 0.012

4.14 299.1 0.008

4.15 298.6 0.564

7

E (eV) k (nm) f

3.15 393.8 0.309

3.35 370.0 0.001

3.97 312.5 0.004

4.05 306.2 0.538

4.23 293.1 0.022

17

E (eV) k (nm) f

2.88 430.4 0.191

3.34 371.8 0.284

3.51 353.6 0.036

3.99 311.0 0.142

4.12 301.1 0.221

8

E (eV) k (nm) f

3.72 333.2 0.069

3.78 327.8 0.113

4.28 289.8 0

4.50 275.6 1.353

4.57 271.2 0

18

E (eV) k (nm) f

3.21 386.6 0.008

3.54 350.1 0.040

3.64 341.0 0.165

4.29 289.0 0

4.45 278.5 0.087

9

E (eV) k (nm) f

3.35 370.3 0.059

3.46 358.4 0.115

3.97 312.1 1.349

4.09 302.9 0

4.21 294.4 0.609

19

E (eV) k (nm) f

4.05 306.3 0.209

4.13 300.3 0.002

4.73 262.0 0.838

4.75 261.0 0.335

5.06 245.3 0.025

10

E (eV) k (nm) f

3.00 414.4 0.750

3.10 399.4 0.009

3.62 342.3 0.108

3.92 316.1 0

4.05 305.8 0.002

a

The sequence numbers of PAHs correspond with those in Fig. 1.

1004

Y. Wang et al. / Chemosphere 76 (2009) 999–1005

coefficient. Several PAHs such as benzo[b]anthracene, dibenzo[a,i]pyrene and indeno[1,2,3-cd]pyrene with small EGAP (2.75–3.30 eV) have absorption of visible light as well as ultraviolet light (Table 4). It was found that dibenzo[a,i]pyrene and indeno[1,2,3-cd]pyrene were highly phototoxic to Daphnia magna or Scenedemus vacuolatus (Table 2). In comparison, fluorene and phenanthrene with the large EGAP (4.94 eV, 4.68 eV) predominately absorb wavelengths less than 290 nm (Table 4). As a result, they exerted the lowest photoinduced toxicity to Daphnia magna or Scenedemus vacuolatus (Table 2). Thus the photoinduced toxicity increased with the decreasing of the EGAP probably because of better spectral overlap between the absorption spectrum of PAHs and the radiation spectrum of light source. The calculated ET1 values of PAHs are much higher than the excited state energy of 1 O2 (1.05 eV) (Llano et al., 2003). It was suggested that the studied 3PAH*s were all able to transfer energy to ground state oxygen (3 O2 ) and generate high reactive 1 O2 through type II photosensitization reaction 3

PAH þ 3 O2 ! 1 PAH þ 1 O2

ð3Þ 3

*

According to TD-DFT calculations, all the studied PAHs may react with 3 O2 to form PAH+ and O 2 (Eq. (4)) because the VIET1 values for PAHs were all lower than adiabatic electron affinity (AEA) for 3 O2 . However, VIET1 was not involved in the developed QSAR models. To some extent, QSAR models results demonstrate that VEAT1 related to PAH production affect the photosensitization type I reaction. The larger the VEAT1, the easier for 3PAH* it is to oxidize the substrate (S) (e.g., amino acids and DNA/RNA bases) nearby to form the PAH and substrate cation (S+) (Eq. (5)). PAH was able to react with 3 O2 to form O 2 through Eq. (6) because the VEAS0 values for PAHs were all estimated to be lower than the AEA for 3 O2 (3.87 eV). Thus, VEAT1 for PAHs plays a key role in type I photosensitization reaction. 3 3

PAH þ 3 O2 ! PAHþ þ O 2 



þ

PAH þ S ! PAH þ S

PAH þ 3 O2 ! PAH þ O 2

ð4Þ ð5Þ ð6Þ

4. Conclusions The present study demonstrates that DFT descriptors can be used to construct QSAR models for photoinduced toxicity of PAHs to aquatic organisms. The average molecular polarizability, which affects the partition behavior, governs the photoinduced toxicity of PAHs. The photochemical properties such as EGAP and ET1 are also significant parameters related to the photoinduced toxicity. The VEAT1 that describes the capability of the PAH generation in photosensitization type I reaction is another molecular structural descriptor involved in the photoinduced toxicity QSAR models. More insight into the photoinduced toxicity mechanism of PAHs was obtained through the current QSAR model results. Acknowledgement The study was supported by the National Basic Research Program of China (2006CB403302). References Arfsten, D.P., Schaeffer, D.J., Mulveny, D.C., 1996. The effects of near ultraviolet radiation on the toxic effects of polycyclic aromatic hydrocarbons in animals and plants: a review. Ecotoxicol. Environ. Saf. 33, 1–24. Betowski, L.D., Enlow, M., Riddick, L., 2002. The phototoxicity of polycyclic aromatic hydrocarbons: a theoretical study of excited states and correlation to experiment. Comput. Chem. 26, 371–377. Chen, J.W., Kong, L.R., Zhu, C.M., Huang, Q.G., Wang, L.S., 1996. Correlation between photolysis rate constants of polycyclic aromatic hydrocarbons and frontier molecular orbital energy. Chemosphere 33, 1143–1150.

Ding, G.H., Chen, J.W., Qiao, X.L., Huang, L.P., Lin, J., Chen, X.Y., 2005. Comparison of subcooled liquid vapor pressures of polychlorinated dibenzo-p-dioxins and dibenzofurans predicted by QSPR and GC-RI methods. SAR QSAR Environ. Res. 16, 301–312. Dong, S., Hwang, H.M., Harrison, C., Holloway, L., Shi, X., Yu, H., 2000. UVA lightinduced DNA cleavage by selected polycyclic aromatic hydrocarbons. Bull. Environ. Contam. Toxicol. 64, 467–474. Dong, S., Fu, P.P., Shirsat, R.N., Hwang, H.M., Leszczynski, J., Yu, H., 2002. UVA lightinduced DNA cleavage by isomeric methylbenz[a]anthracenes. Chem. Res. Toxicol. 15, 400–407. El-Alawi, Y.S., Huang, X.D., Dixon, D.G., Greenberg, B.M., 2002. Quantitative structure-activity relationship for the photoinduced toxicity of polycyclic aromatic hydrocarbons to the luminescent bacteria Vibrio fischeri. Environ. Toxicol. Chem. 21, 2225–2232. Enoch, S.J., Cronin, M.T., Schultz, T.W., Madden, J.C., 2008. Quantitative and mechanistic read across for predicting the skin sensitization potential of alkenes acting via Michael addition. Chem. Res. Toxicol. 21, 513–520. Escher, B.I., Hermens, J.L.M., 2002. Modes of action in ecotoxicology: their role in body burdens, species sensitivity, QSARs, and mixture effects. Environ. Sci. Technol. 36, 4201–4217. Fasnacht, M.P., Blough, N.V., 2003. Mechanisms of the aqueous photodegradation of polycyclic aromatic hydrocarbons. Environ. Sci. Technol. 37, 5767–5772. Foote, C.S., 1991. Definition of type I and type II photosensitized oxidation. Photochem. Photobiol. 54, 659. Frisch, M.J., Trucks, G.W., Schlegel, H.B., Scuseria, G.E., Robb, M.A., Cheeseman, J.R., Montgomery Jr., J.A., Vreven, T., Kudin, K.N., Burant, J.C., Millam, J.M., Iyengar, S.S., Tomasi, J., Barone, V., Mennucci, B., Cossi, M., Scalmani, G., Rega, N., Petersson, G.A., Nakatsuji, H., Hada, M., Ehara, M., Toyota, K., Fukuda, R., Hasegawa, J., Ishida, M., Nakajima, T., Honda, Y.K., Kitao, O., Nakai, H., Klene, M., Li, X., Knox, J.E., Hratchian, H.P., Cross, J.B., Bakken, V., Adamo, C., Jaramillo, J., Gomperts, R., Stratmann, R.E., Yazyev, O., Austin, A.J., Cammi, R., Pomelli, C., Ochterski, J.W., Ayala, P.Y., Morokuma, K., Voth, G.A., Salvador, P., Dannenberg, J.J., Zakrzewski, V.G., Dapprich, S., Daniels, A.D., Strain, M.C., Farkas, O., Malick, D.K., Rabuck, A.D., Raghavachari, K., Foresman, J.B., Ortiz, J.V., Cui, Q., Baboul, A.G., Clifford, S., Cioslowski, J., Stefanov, B.B., Liu, G., Liashenko, A., Piskorz, P., Komaromi, I., Martin, R.L., Fox, D.J., Keith, T., Al-Laham, M.A., Peng, C.Y., Nanayakkara, A., Challacombe, M., Gill, P.M.W., Johnson, B., Chen, W., Wong, M.W., Gonzalez, C., Pople, J.A., 2003. Gaussian 03. Gaussian Inc., Pittsburgh, PA, USA. Grote, M., Schüürmann, G., Altenburger, R., 2005. Modeling photoinduced algal toxicity of polycyclic aromatic hydrocarbons. Environ. Sci. Technol. 39, 4141– 4149. Huang, X.D., Dixon, D.G., Greenberg, B.M., 1993. Impacts of UV radiation and photomodification on the toxicity of PAHs to the higher plant Lemna gibba (duckweed). Environ. Toxicol. Chem. 12, 1067–1077. Huang, X.D., Dixon, D.G., Greenberg, B.M., 1995. Increased polycyclic aromatic hydrocarbon toxicity following their photomodification in natural sunlight— impacts on the duckweed Lemna-gibba L. G-3. Ecotoxicol. Environ. Saf. 32, 194– 200. Huang, X.D., Krylov, S.N., Ren, L.S., McConkey, B.J., Dixon, D.G., Greenberg, B.M., 1997. Mechanistic quantitative structure–activity relationship model for the photoinduced toxicity of polycyclic aromatic hydrocarbons. II. An empirical model for the toxicity of 16 polycyclic aromatic hydrocarbons to the duckweed Lemna gibba L. G-3. Environ. Toxicol. Chem. 16, 2296–2303. Huovinen, P.S., Soimasuo, M.R., Oikari, A.O.J., 2001. Photoinduced toxicity of retene to Daphnia magna under enhanced UV-B radiation. Chemosphere 45, 683–691. Johnson, L.R., Davenport, R., Balbach, H., Schaeffer, D.J., 1994. Phototoxicology. 3. Comparative toxicity of trinitrotoluene and aminodinitrotoluenes to Daphnia magna, Dugesia dorotocephala, and sheep erythrocytes. Ecotoxicol. Environ. Saf. 27, 34–49. Karelson, M., Lobanov, V.S., Katritzky, A.R., 1996. Quantum-chemical descriptors in QSAR/QSPR studies. Chem. Rev. 96, 1027–1044. Lampi, M.A., Gurska, J., McDonald, K.I.C., Xie, F.L., Huang, X.D., Dixon, D.G., Greenberg, B.M., 2005. Photoinduced toxicity of polycyclic aromatic hydrocarbons to Daphnia magna: ultraviolet-mediated effects and the toxicity of polycyclic aromatic hydrocarbon photoproducts. Environ. Toxicol. Chem. 25, 1079–1087. Lampi, M.A., Gurska, J., Huang, X.D., Dixon, D.G., Greenberg, B.M., 2007. A predictive quantitative structure–activity relationship model for the photoinduced toxicity of polycyclic aromatic hydrocarbons to Daphnia magna with the use of factors for photosensitization and photomodification. Environ. Toxicol. Chem. 26, 406–415. Little, E.E., Cleveland, L., Calfee, R., Barron, M.G., 2000. Assessment of the photoenhanced toxicity of a weathered oil to the tidewater silverside. Environ. Toxicol. Chem. 19, 926–932. Llano, J., Raber, J., Eriksson, L.A., 2003. Theoretical study of phototoxic reactions of psoralens. J. Photochem. Photobiol. A: Chem. 154, 235–243. McConkey, B.J., Duxbury, C.L., Dixon, D.G., Greenberg, B.M., 1997. Toxicity of a PAH photooxidation product to the bacteria Photobacterium phosphoreum and the duckweed Lemna gibba: effects of phenanthrene and its primary photoproduct Phenanthrenequinone. Environ. Toxicol. Chem. 16, 892–899. Mekenyan, O.G., Ankley, G.T., Veith, G.D., Call, D.J., 1994. QSARs for photoinduced toxicity: I acute lethality of polycyclic aromatic hydrocarbons to Daphnia magna. Chemosphere 28, 567–582. Musa, K.A., Eriksson, L.A., 2007. Theoretical study of ibuprofen phototoxicity. J. Phys. Chem. B 111, 13345–13352.

Y. Wang et al. / Chemosphere 76 (2009) 999–1005 Newsted, J.L., Giesy, J.P., 1987. Predictive models of photoinduced acute toxicity of polycyclic aromatic hydrocarbons to Daphnia magna, strauss (Cladoera, Crustacea). Environ. Toxicol. Chem. 6, 445–461. Niu, J.F., Yang, Z.F., Shen, Z.Y., Wang, L.L., 2006. QSPRs for the prediction of photodegradation half-life of PCBs in n-hexane. SAR QSAR Environ. Res. 17, 173–182. Oris, J.T., Giesy, J.P., 1985. The photoenhanced toxicity of anthracene to juvenile sunfish (Lepomis spp.). Aquat. Toxicol 6, 133–146. Pasha, F.A., Muddassar, M., Beg, Y., Cho, S.J., 2008. DFT-based de novo QSAR of Phenoloxidase Inhibitors. Chem. Biol. Drug Des. 71, 483–493. Shen, L., Ji, H.F., Zhang, H.Y., 2007. Computational note on the photosensitization mechanisms of riboflavin. J. Mol. Struct. (Theochem) 821, 171–172. Svenson, A., Hynning, P.A., 1997. Increased aquatic toxicity following photolytic conversion of an organochlorine pollutant. Chemosphere 34, 1685–1692. Toyooka, T., Ibuki, Y., 2007. DNA damage induced by coexposure to PAHs and light. Environ. Toxicol. Pharmacol. 23, 256–263.

1005

Veith, G.D., Mekenyan, O.G., Ankley, G.T., Call, D.J., 1995a. A QSAR analysis of substituent effects on the photoinduced acute toxicity of PAHs. Chemosphere 30, 2129–2142. Veith, G.D., Mekenyan, O.G., Ankley, G.T., Call, D.J., 1995b. Qsar evaluation of alphaterthienyl phototoxicity. Environ. Sci. Technol. 29, 1267–1272. Wang, S., Sheng, Y., Feng, M., Leszczynski, J., Wang, L., Tachikawa, H., Yu, H., 2007. Light-induced cytotoxicity of 16 polycyclic aromatic hydrocarbons on the US EPA priority pollutant list in human skin HaCaT keratinocytes: relationship between phototoxicity and excited state properties. Environ. Toxicol. 22, 318– 327. Wiegman, S., Van Vlaardingen, P.L.A., Peijnenburg, W.J.G.M., Van Beusekom, S.A.M., Kraak, M.H.S., Admiraal, W., 1999. Photokinetics of azaarenes and toxicity of phototransformation products to the marine diatom Phaeodactylum tricornutum. Environ. Sci. Technol. 33, 4256–4262.